Skip to main content
. 2008 Dec 5;19(7-8):631–641. doi: 10.1080/10629360802550473

Table 2.

Cross-validation results for the QSAR models.

Sensitivity (%) Specificity (%) Concordance (%)
Reproductive toxicity
 Teratogenic risk 50.2 91.3 79.3
 Rodent dominant lethal 41.3 95.2 75.9
Drosophila m. SLRL 73.9 88.1 81.6
Endocrine disruption
 Estrogen αreceptor binding 77.3 86.5 82.5
 Estrogen reporter gene 46.4 94.9 80.9
 Androgen receptor antagonism* 64.4 84.2 76.1

Note: The androgen receptor antagonism model was also validated by prospective external validation, where a validation set was chosen after the model had been constructed by randomly choosing test chemicals within the applicability domain. The chemicals were double blinded and tested in our laboratory. The external validation with 102 chemicals resulted in a sensitivity of 57%, a specificity of 98%, and a concordance of 92% of the model [14].